1 /*M/////////////////////////////////////////////////////////////////////////////////////// 2 // 3 // IMPORTANT: READ BEFORE DOWNLOADING, COPYING, INSTALLING OR USING. 4 // 5 // By downloading, copying, installing or using the software you agree to this license. 6 // If you do not agree to this license, do not download, install, 7 // copy or use the software. 8 // 9 // 10 // License Agreement 11 // For Open Source Computer Vision Library 12 // 13 // Copyright (C) 2000-2008, Intel Corporation, all rights reserved. 14 // Copyright (C) 2009, Willow Garage Inc., all rights reserved. 15 // Third party copyrights are property of their respective owners. 16 // 17 // Redistribution and use in source and binary forms, with or without modification, 18 // are permitted provided that the following conditions are met: 19 // 20 // * Redistribution's of source code must retain the above copyright notice, 21 // this list of conditions and the following disclaimer. 22 // 23 // * Redistribution's in binary form must reproduce the above copyright notice, 24 // this list of conditions and the following disclaimer in the documentation 25 // and/or other materials provided with the distribution. 26 // 27 // * The name of the copyright holders may not be used to endorse or promote products 28 // derived from this software without specific prior written permission. 29 // 30 // This software is provided by the copyright holders and contributors "as is" and 31 // any express or implied warranties, including, but not limited to, the implied 32 // warranties of merchantability and fitness for a particular purpose are disclaimed. 33 // In no event shall the Intel Corporation or contributors be liable for any direct, 34 // indirect, incidental, special, exemplary, or consequential damages 35 // (including, but not limited to, procurement of substitute goods or services; 36 // loss of use, data, or profits; or business interruption) however caused 37 // and on any theory of liability, whether in contract, strict liability, 38 // or tort (including negligence or otherwise) arising in any way out of 39 // the use of this software, even if advised of the possibility of such damage. 40 // 41 //M*/ 42 43 #include "test_precomp.hpp" 44 #include <time.h> 45 46 #define CALIB3D_HOMOGRAPHY_ERROR_MATRIX_SIZE 1 47 #define CALIB3D_HOMOGRAPHY_ERROR_MATRIX_DIFF 2 48 #define CALIB3D_HOMOGRAPHY_ERROR_REPROJ_DIFF 3 49 #define CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK 4 50 #define CALIB3D_HOMOGRAPHY_ERROR_RANSAC_DIFF 5 51 52 #define MESSAGE_MATRIX_SIZE "Homography matrix must have 3*3 sizes." 53 #define MESSAGE_MATRIX_DIFF "Accuracy of homography transformation matrix less than required." 54 #define MESSAGE_REPROJ_DIFF_1 "Reprojection error for current pair of points more than required." 55 #define MESSAGE_REPROJ_DIFF_2 "Reprojection error is not optimal." 56 #define MESSAGE_RANSAC_MASK_1 "Sizes of inliers/outliers mask are incorrect." 57 #define MESSAGE_RANSAC_MASK_2 "Mask mustn't have any outliers." 58 #define MESSAGE_RANSAC_MASK_3 "All values of mask must be 1 (true) or 0 (false)." 59 #define MESSAGE_RANSAC_MASK_4 "Mask of inliers/outliers is incorrect." 60 #define MESSAGE_RANSAC_MASK_5 "Inlier in original mask shouldn't be outlier in found mask." 61 #define MESSAGE_RANSAC_DIFF "Reprojection error for current pair of points more than required." 62 63 #define MAX_COUNT_OF_POINTS 303 64 #define COUNT_NORM_TYPES 3 65 #define METHODS_COUNT 4 66 67 int NORM_TYPE[COUNT_NORM_TYPES] = {cv::NORM_L1, cv::NORM_L2, cv::NORM_INF}; 68 int METHOD[METHODS_COUNT] = {0, cv::RANSAC, cv::LMEDS, cv::RHO}; 69 70 using namespace cv; 71 using namespace std; 72 73 class CV_HomographyTest: public cvtest::ArrayTest 74 { 75 public: 76 CV_HomographyTest(); 77 ~CV_HomographyTest(); 78 79 void run (int); 80 81 protected: 82 83 int method; 84 int image_size; 85 double reproj_threshold; 86 double sigma; 87 88 private: 89 float max_diff, max_2diff; 90 bool check_matrix_size(const cv::Mat& H); 91 bool check_matrix_diff(const cv::Mat& original, const cv::Mat& found, const int norm_type, double &diff); 92 int check_ransac_mask_1(const Mat& src, const Mat& mask); 93 int check_ransac_mask_2(const Mat& original_mask, const Mat& found_mask); 94 95 void print_information_1(int j, int N, int method, const Mat& H); 96 void print_information_2(int j, int N, int method, const Mat& H, const Mat& H_res, int k, double diff); 97 void print_information_3(int method, int j, int N, const Mat& mask); 98 void print_information_4(int method, int j, int N, int k, int l, double diff); 99 void print_information_5(int method, int j, int N, int l, double diff); 100 void print_information_6(int method, int j, int N, int k, double diff, bool value); 101 void print_information_7(int method, int j, int N, int k, double diff, bool original_value, bool found_value); 102 void print_information_8(int method, int j, int N, int k, int l, double diff); 103 }; 104 105 CV_HomographyTest::CV_HomographyTest() : max_diff(1e-2f), max_2diff(2e-2f) 106 { 107 method = 0; 108 image_size = 100; 109 reproj_threshold = 3.0; 110 sigma = 0.01; 111 } 112 113 CV_HomographyTest::~CV_HomographyTest() {} 114 115 bool CV_HomographyTest::check_matrix_size(const cv::Mat& H) 116 { 117 return (H.rows == 3) && (H.cols == 3); 118 } 119 120 bool CV_HomographyTest::check_matrix_diff(const cv::Mat& original, const cv::Mat& found, const int norm_type, double &diff) 121 { 122 diff = cvtest::norm(original, found, norm_type); 123 return diff <= max_diff; 124 } 125 126 int CV_HomographyTest::check_ransac_mask_1(const Mat& src, const Mat& mask) 127 { 128 if (!(mask.cols == 1) && (mask.rows == src.cols)) return 1; 129 if (countNonZero(mask) < mask.rows) return 2; 130 for (int i = 0; i < mask.rows; ++i) if (mask.at<uchar>(i, 0) > 1) return 3; 131 return 0; 132 } 133 134 int CV_HomographyTest::check_ransac_mask_2(const Mat& original_mask, const Mat& found_mask) 135 { 136 if (!(found_mask.cols == 1) && (found_mask.rows == original_mask.rows)) return 1; 137 for (int i = 0; i < found_mask.rows; ++i) if (found_mask.at<uchar>(i, 0) > 1) return 2; 138 return 0; 139 } 140 141 void CV_HomographyTest::print_information_1(int j, int N, int _method, const Mat& H) 142 { 143 cout << endl; cout << "Checking for homography matrix sizes..." << endl; cout << endl; 144 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 145 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 146 cout << "Count of points: " << N << endl; cout << endl; 147 cout << "Method: "; if (_method == 0) cout << 0; else if (_method == 8) cout << "RANSAC"; else if (_method == cv::RHO) cout << "RHO"; else cout << "LMEDS"; cout << endl; 148 cout << "Homography matrix:" << endl; cout << endl; 149 cout << H << endl; cout << endl; 150 cout << "Number of rows: " << H.rows << " Number of cols: " << H.cols << endl; cout << endl; 151 } 152 153 void CV_HomographyTest::print_information_2(int j, int N, int _method, const Mat& H, const Mat& H_res, int k, double diff) 154 { 155 cout << endl; cout << "Checking for accuracy of homography matrix computing..." << endl; cout << endl; 156 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 157 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 158 cout << "Count of points: " << N << endl; cout << endl; 159 cout << "Method: "; if (_method == 0) cout << 0; else if (_method == 8) cout << "RANSAC"; else if (_method == cv::RHO) cout << "RHO"; else cout << "LMEDS"; cout << endl; 160 cout << "Original matrix:" << endl; cout << endl; 161 cout << H << endl; cout << endl; 162 cout << "Found matrix:" << endl; cout << endl; 163 cout << H_res << endl; cout << endl; 164 cout << "Norm type using in criteria: "; if (NORM_TYPE[k] == 1) cout << "INF"; else if (NORM_TYPE[k] == 2) cout << "L1"; else cout << "L2"; cout << endl; 165 cout << "Difference between matrices: " << diff << endl; 166 cout << "Maximum allowed difference: " << max_diff << endl; cout << endl; 167 } 168 169 void CV_HomographyTest::print_information_3(int _method, int j, int N, const Mat& mask) 170 { 171 cout << endl; cout << "Checking for inliers/outliers mask..." << endl; cout << endl; 172 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 173 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 174 cout << "Count of points: " << N << endl; cout << endl; 175 cout << "Method: "; if (_method == RANSAC) cout << "RANSAC" << endl; else if (_method == cv::RHO) cout << "RHO" << endl; else cout << _method << endl; 176 cout << "Found mask:" << endl; cout << endl; 177 cout << mask << endl; cout << endl; 178 cout << "Number of rows: " << mask.rows << " Number of cols: " << mask.cols << endl; cout << endl; 179 } 180 181 void CV_HomographyTest::print_information_4(int _method, int j, int N, int k, int l, double diff) 182 { 183 cout << endl; cout << "Checking for accuracy of reprojection error computing..." << endl; cout << endl; 184 cout << "Method: "; if (_method == 0) cout << 0 << endl; else cout << "CV_LMEDS" << endl; 185 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 186 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 187 cout << "Sigma of normal noise: " << sigma << endl; 188 cout << "Count of points: " << N << endl; 189 cout << "Number of point: " << k << endl; 190 cout << "Norm type using in criteria: "; if (NORM_TYPE[l] == 1) cout << "INF"; else if (NORM_TYPE[l] == 2) cout << "L1"; else cout << "L2"; cout << endl; 191 cout << "Difference with noise of point: " << diff << endl; 192 cout << "Maxumum allowed difference: " << max_2diff << endl; cout << endl; 193 } 194 195 void CV_HomographyTest::print_information_5(int _method, int j, int N, int l, double diff) 196 { 197 cout << endl; cout << "Checking for accuracy of reprojection error computing..." << endl; cout << endl; 198 cout << "Method: "; if (_method == 0) cout << 0 << endl; else cout << "CV_LMEDS" << endl; 199 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 200 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 201 cout << "Sigma of normal noise: " << sigma << endl; 202 cout << "Count of points: " << N << endl; 203 cout << "Norm type using in criteria: "; if (NORM_TYPE[l] == 1) cout << "INF"; else if (NORM_TYPE[l] == 2) cout << "L1"; else cout << "L2"; cout << endl; 204 cout << "Difference with noise of points: " << diff << endl; 205 cout << "Maxumum allowed difference: " << max_diff << endl; cout << endl; 206 } 207 208 void CV_HomographyTest::print_information_6(int _method, int j, int N, int k, double diff, bool value) 209 { 210 cout << endl; cout << "Checking for inliers/outliers mask..." << endl; cout << endl; 211 cout << "Method: "; if (_method == RANSAC) cout << "RANSAC" << endl; else if (_method == cv::RHO) cout << "RHO" << endl; else cout << _method << endl; 212 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 213 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 214 cout << "Count of points: " << N << " " << endl; 215 cout << "Number of point: " << k << " " << endl; 216 cout << "Reprojection error for this point: " << diff << " " << endl; 217 cout << "Reprojection error threshold: " << reproj_threshold << " " << endl; 218 cout << "Value of found mask: "<< value << endl; cout << endl; 219 } 220 221 void CV_HomographyTest::print_information_7(int _method, int j, int N, int k, double diff, bool original_value, bool found_value) 222 { 223 cout << endl; cout << "Checking for inliers/outliers mask..." << endl; cout << endl; 224 cout << "Method: "; if (_method == RANSAC) cout << "RANSAC" << endl; else if (_method == cv::RHO) cout << "RHO" << endl; else cout << _method << endl; 225 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 226 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 227 cout << "Count of points: " << N << " " << endl; 228 cout << "Number of point: " << k << " " << endl; 229 cout << "Reprojection error for this point: " << diff << " " << endl; 230 cout << "Reprojection error threshold: " << reproj_threshold << " " << endl; 231 cout << "Value of original mask: "<< original_value << " Value of found mask: " << found_value << endl; cout << endl; 232 } 233 234 void CV_HomographyTest::print_information_8(int _method, int j, int N, int k, int l, double diff) 235 { 236 cout << endl; cout << "Checking for reprojection error of inlier..." << endl; cout << endl; 237 cout << "Method: "; if (_method == RANSAC) cout << "RANSAC" << endl; else if (_method == cv::RHO) cout << "RHO" << endl; else cout << _method << endl; 238 cout << "Sigma of normal noise: " << sigma << endl; 239 cout << "Type of srcPoints: "; if ((j>-1) && (j<2)) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; 240 cout << " Type of dstPoints: "; if (j % 2 == 0) cout << "Mat of CV_32FC2"; else cout << "vector <Point2f>"; cout << endl; 241 cout << "Count of points: " << N << " " << endl; 242 cout << "Number of point: " << k << " " << endl; 243 cout << "Norm type using in criteria: "; if (NORM_TYPE[l] == 1) cout << "INF"; else if (NORM_TYPE[l] == 2) cout << "L1"; else cout << "L2"; cout << endl; 244 cout << "Difference with noise of point: " << diff << endl; 245 cout << "Maxumum allowed difference: " << max_2diff << endl; cout << endl; 246 } 247 248 void CV_HomographyTest::run(int) 249 { 250 for (int N = 4; N <= MAX_COUNT_OF_POINTS; ++N) 251 { 252 RNG& rng = ts->get_rng(); 253 254 float *src_data = new float [2*N]; 255 256 for (int i = 0; i < N; ++i) 257 { 258 src_data[2*i] = (float)cvtest::randReal(rng)*image_size; 259 src_data[2*i+1] = (float)cvtest::randReal(rng)*image_size; 260 } 261 262 cv::Mat src_mat_2f(1, N, CV_32FC2, src_data), 263 src_mat_2d(2, N, CV_32F, src_data), 264 src_mat_3d(3, N, CV_32F); 265 cv::Mat dst_mat_2f, dst_mat_2d, dst_mat_3d; 266 267 vector <Point2f> src_vec, dst_vec; 268 269 for (int i = 0; i < N; ++i) 270 { 271 float *tmp = src_mat_2d.ptr<float>()+2*i; 272 src_mat_3d.at<float>(0, i) = tmp[0]; 273 src_mat_3d.at<float>(1, i) = tmp[1]; 274 src_mat_3d.at<float>(2, i) = 1.0f; 275 276 src_vec.push_back(Point2f(tmp[0], tmp[1])); 277 } 278 279 double fi = cvtest::randReal(rng)*2*CV_PI; 280 281 double t_x = cvtest::randReal(rng)*sqrt(image_size*1.0), 282 t_y = cvtest::randReal(rng)*sqrt(image_size*1.0); 283 284 double Hdata[9] = { cos(fi), -sin(fi), t_x, 285 sin(fi), cos(fi), t_y, 286 0.0f, 0.0f, 1.0f }; 287 288 cv::Mat H_64(3, 3, CV_64F, Hdata), H_32; 289 290 H_64.convertTo(H_32, CV_32F); 291 292 dst_mat_3d = H_32*src_mat_3d; 293 294 dst_mat_2d.create(2, N, CV_32F); dst_mat_2f.create(1, N, CV_32FC2); 295 296 for (int i = 0; i < N; ++i) 297 { 298 float *tmp_2f = dst_mat_2f.ptr<float>()+2*i; 299 tmp_2f[0] = dst_mat_2d.at<float>(0, i) = dst_mat_3d.at<float>(0, i) /= dst_mat_3d.at<float>(2, i); 300 tmp_2f[1] = dst_mat_2d.at<float>(1, i) = dst_mat_3d.at<float>(1, i) /= dst_mat_3d.at<float>(2, i); 301 dst_mat_3d.at<float>(2, i) = 1.0f; 302 303 dst_vec.push_back(Point2f(tmp_2f[0], tmp_2f[1])); 304 } 305 306 for (int i = 0; i < METHODS_COUNT; ++i) 307 { 308 method = METHOD[i]; 309 switch (method) 310 { 311 case 0: 312 case LMEDS: 313 { 314 Mat H_res_64 [4] = { cv::findHomography(src_mat_2f, dst_mat_2f, method), 315 cv::findHomography(src_mat_2f, dst_vec, method), 316 cv::findHomography(src_vec, dst_mat_2f, method), 317 cv::findHomography(src_vec, dst_vec, method) }; 318 319 for (int j = 0; j < 4; ++j) 320 { 321 322 if (!check_matrix_size(H_res_64[j])) 323 { 324 print_information_1(j, N, method, H_res_64[j]); 325 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_MATRIX_SIZE, MESSAGE_MATRIX_SIZE); 326 return; 327 } 328 329 double diff; 330 331 for (int k = 0; k < COUNT_NORM_TYPES; ++k) 332 if (!check_matrix_diff(H_64, H_res_64[j], NORM_TYPE[k], diff)) 333 { 334 print_information_2(j, N, method, H_64, H_res_64[j], k, diff); 335 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_MATRIX_DIFF, MESSAGE_MATRIX_DIFF); 336 return; 337 } 338 } 339 340 continue; 341 } 342 case cv::RHO: 343 case RANSAC: 344 { 345 cv::Mat mask [4]; double diff; 346 347 Mat H_res_64 [4] = { cv::findHomography(src_mat_2f, dst_mat_2f, method, reproj_threshold, mask[0]), 348 cv::findHomography(src_mat_2f, dst_vec, method, reproj_threshold, mask[1]), 349 cv::findHomography(src_vec, dst_mat_2f, method, reproj_threshold, mask[2]), 350 cv::findHomography(src_vec, dst_vec, method, reproj_threshold, mask[3]) }; 351 352 for (int j = 0; j < 4; ++j) 353 { 354 355 if (!check_matrix_size(H_res_64[j])) 356 { 357 print_information_1(j, N, method, H_res_64[j]); 358 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_MATRIX_SIZE, MESSAGE_MATRIX_SIZE); 359 return; 360 } 361 362 for (int k = 0; k < COUNT_NORM_TYPES; ++k) 363 if (!check_matrix_diff(H_64, H_res_64[j], NORM_TYPE[k], diff)) 364 { 365 print_information_2(j, N, method, H_64, H_res_64[j], k, diff); 366 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_MATRIX_DIFF, MESSAGE_MATRIX_DIFF); 367 return; 368 } 369 370 int code = check_ransac_mask_1(src_mat_2f, mask[j]); 371 372 if (code) 373 { 374 print_information_3(method, j, N, mask[j]); 375 376 switch (code) 377 { 378 case 1: { CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK, MESSAGE_RANSAC_MASK_1); break; } 379 case 2: { CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK, MESSAGE_RANSAC_MASK_2); break; } 380 case 3: { CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK, MESSAGE_RANSAC_MASK_3); break; } 381 382 default: break; 383 } 384 385 return; 386 } 387 388 } 389 390 continue; 391 } 392 393 default: continue; 394 } 395 } 396 397 Mat noise_2f(1, N, CV_32FC2); 398 rng.fill(noise_2f, RNG::NORMAL, Scalar::all(0), Scalar::all(sigma)); 399 400 cv::Mat mask(N, 1, CV_8UC1); 401 402 for (int i = 0; i < N; ++i) 403 { 404 float *a = noise_2f.ptr<float>()+2*i, *_2f = dst_mat_2f.ptr<float>()+2*i; 405 _2f[0] += a[0]; _2f[1] += a[1]; 406 mask.at<bool>(i, 0) = !(sqrt(a[0]*a[0]+a[1]*a[1]) > reproj_threshold); 407 } 408 409 for (int i = 0; i < METHODS_COUNT; ++i) 410 { 411 method = METHOD[i]; 412 switch (method) 413 { 414 case 0: 415 case LMEDS: 416 { 417 Mat H_res_64 [4] = { cv::findHomography(src_mat_2f, dst_mat_2f), 418 cv::findHomography(src_mat_2f, dst_vec), 419 cv::findHomography(src_vec, dst_mat_2f), 420 cv::findHomography(src_vec, dst_vec) }; 421 422 for (int j = 0; j < 4; ++j) 423 { 424 425 if (!check_matrix_size(H_res_64[j])) 426 { 427 print_information_1(j, N, method, H_res_64[j]); 428 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_MATRIX_SIZE, MESSAGE_MATRIX_SIZE); 429 return; 430 } 431 432 Mat H_res_32; H_res_64[j].convertTo(H_res_32, CV_32F); 433 434 cv::Mat dst_res_3d(3, N, CV_32F), noise_2d(2, N, CV_32F); 435 436 for (int k = 0; k < N; ++k) 437 { 438 439 Mat tmp_mat_3d = H_res_32*src_mat_3d.col(k); 440 441 dst_res_3d.at<float>(0, k) = tmp_mat_3d.at<float>(0, 0) /= tmp_mat_3d.at<float>(2, 0); 442 dst_res_3d.at<float>(1, k) = tmp_mat_3d.at<float>(1, 0) /= tmp_mat_3d.at<float>(2, 0); 443 dst_res_3d.at<float>(2, k) = tmp_mat_3d.at<float>(2, 0) = 1.0f; 444 445 float *a = noise_2f.ptr<float>()+2*k; 446 noise_2d.at<float>(0, k) = a[0]; noise_2d.at<float>(1, k) = a[1]; 447 448 for (int l = 0; l < COUNT_NORM_TYPES; ++l) 449 if (cv::norm(tmp_mat_3d, dst_mat_3d.col(k), NORM_TYPE[l]) - cv::norm(noise_2d.col(k), NORM_TYPE[l]) > max_2diff) 450 { 451 print_information_4(method, j, N, k, l, cv::norm(tmp_mat_3d, dst_mat_3d.col(k), NORM_TYPE[l]) - cv::norm(noise_2d.col(k), NORM_TYPE[l])); 452 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_REPROJ_DIFF, MESSAGE_REPROJ_DIFF_1); 453 return; 454 } 455 456 } 457 458 for (int l = 0; l < COUNT_NORM_TYPES; ++l) 459 if (cv::norm(dst_res_3d, dst_mat_3d, NORM_TYPE[l]) - cv::norm(noise_2d, NORM_TYPE[l]) > max_diff) 460 { 461 print_information_5(method, j, N, l, cv::norm(dst_res_3d, dst_mat_3d, NORM_TYPE[l]) - cv::norm(noise_2d, NORM_TYPE[l])); 462 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_REPROJ_DIFF, MESSAGE_REPROJ_DIFF_2); 463 return; 464 } 465 466 } 467 468 continue; 469 } 470 case cv::RHO: 471 case RANSAC: 472 { 473 cv::Mat mask_res [4]; 474 475 Mat H_res_64 [4] = { cv::findHomography(src_mat_2f, dst_mat_2f, method, reproj_threshold, mask_res[0]), 476 cv::findHomography(src_mat_2f, dst_vec, method, reproj_threshold, mask_res[1]), 477 cv::findHomography(src_vec, dst_mat_2f, method, reproj_threshold, mask_res[2]), 478 cv::findHomography(src_vec, dst_vec, method, reproj_threshold, mask_res[3]) }; 479 480 for (int j = 0; j < 4; ++j) 481 { 482 if (!check_matrix_size(H_res_64[j])) 483 { 484 print_information_1(j, N, method, H_res_64[j]); 485 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_MATRIX_SIZE, MESSAGE_MATRIX_SIZE); 486 return; 487 } 488 489 int code = check_ransac_mask_2(mask, mask_res[j]); 490 491 if (code) 492 { 493 print_information_3(method, j, N, mask_res[j]); 494 495 switch (code) 496 { 497 case 1: { CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK, MESSAGE_RANSAC_MASK_1); break; } 498 case 2: { CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK, MESSAGE_RANSAC_MASK_3); break; } 499 500 default: break; 501 } 502 503 return; 504 } 505 506 cv::Mat H_res_32; H_res_64[j].convertTo(H_res_32, CV_32F); 507 508 cv::Mat dst_res_3d = H_res_32*src_mat_3d; 509 510 for (int k = 0; k < N; ++k) 511 { 512 dst_res_3d.at<float>(0, k) /= dst_res_3d.at<float>(2, k); 513 dst_res_3d.at<float>(1, k) /= dst_res_3d.at<float>(2, k); 514 dst_res_3d.at<float>(2, k) = 1.0f; 515 516 float *p = dst_mat_2f.ptr<float>()+2*k; 517 518 dst_mat_3d.at<float>(0, k) = p[0]; 519 dst_mat_3d.at<float>(1, k) = p[1]; 520 521 double diff = cv::norm(dst_res_3d.col(k), dst_mat_3d.col(k), NORM_L2); 522 523 if (mask_res[j].at<bool>(k, 0) != (diff <= reproj_threshold)) 524 { 525 print_information_6(method, j, N, k, diff, mask_res[j].at<bool>(k, 0)); 526 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK, MESSAGE_RANSAC_MASK_4); 527 return; 528 } 529 530 if (mask.at<bool>(k, 0) && !mask_res[j].at<bool>(k, 0)) 531 { 532 print_information_7(method, j, N, k, diff, mask.at<bool>(k, 0), mask_res[j].at<bool>(k, 0)); 533 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_MASK, MESSAGE_RANSAC_MASK_5); 534 return; 535 } 536 537 if (mask_res[j].at<bool>(k, 0)) 538 { 539 float *a = noise_2f.ptr<float>()+2*k; 540 dst_mat_3d.at<float>(0, k) -= a[0]; 541 dst_mat_3d.at<float>(1, k) -= a[1]; 542 543 cv::Mat noise_2d(2, 1, CV_32F); 544 noise_2d.at<float>(0, 0) = a[0]; noise_2d.at<float>(1, 0) = a[1]; 545 546 for (int l = 0; l < COUNT_NORM_TYPES; ++l) 547 { 548 diff = cv::norm(dst_res_3d.col(k), dst_mat_3d.col(k), NORM_TYPE[l]); 549 550 if (diff - cv::norm(noise_2d, NORM_TYPE[l]) > max_2diff) 551 { 552 print_information_8(method, j, N, k, l, diff - cv::norm(noise_2d, NORM_TYPE[l])); 553 CV_Error(CALIB3D_HOMOGRAPHY_ERROR_RANSAC_DIFF, MESSAGE_RANSAC_DIFF); 554 return; 555 } 556 } 557 } 558 } 559 } 560 561 continue; 562 } 563 564 default: continue; 565 } 566 } 567 } 568 } 569 570 TEST(Calib3d_Homography, accuracy) { CV_HomographyTest test; test.safe_run(); } 571 572 TEST(Calib3d_Homography, EKcase) 573 { 574 float pt1data[] = 575 { 576 2.80073029e+002f, 2.39591217e+002f, 2.21912201e+002f, 2.59783997e+002f, 577 2.16053192e+002f, 2.78826569e+002f, 2.22782532e+002f, 2.82330383e+002f, 578 2.09924820e+002f, 2.89122559e+002f, 2.11077698e+002f, 2.89384674e+002f, 579 2.25287689e+002f, 2.88795532e+002f, 2.11180801e+002f, 2.89653503e+002f, 580 2.24126404e+002f, 2.90466064e+002f, 2.10914429e+002f, 2.90886963e+002f, 581 2.23439362e+002f, 2.91657715e+002f, 2.24809387e+002f, 2.91891602e+002f, 582 2.09809082e+002f, 2.92891113e+002f, 2.08771164e+002f, 2.93093231e+002f, 583 2.23160095e+002f, 2.93259460e+002f, 2.07874023e+002f, 2.93989990e+002f, 584 2.08963638e+002f, 2.94209839e+002f, 2.23963165e+002f, 2.94479645e+002f, 585 2.23241791e+002f, 2.94887817e+002f, 2.09438782e+002f, 2.95233337e+002f, 586 2.08901886e+002f, 2.95762878e+002f, 2.21867981e+002f, 2.95747711e+002f, 587 2.24195511e+002f, 2.98270905e+002f, 2.09331345e+002f, 3.05958191e+002f, 588 2.24727875e+002f, 3.07186035e+002f, 2.26718842e+002f, 3.08095795e+002f, 589 2.25363953e+002f, 3.08200226e+002f, 2.19897797e+002f, 3.13845093e+002f, 590 2.25013474e+002f, 3.15558777e+002f 591 }; 592 593 float pt2data[] = 594 { 595 1.84072723e+002f, 1.43591202e+002f, 1.25912483e+002f, 1.63783859e+002f, 596 2.06439407e+002f, 2.20573929e+002f, 1.43801437e+002f, 1.80703903e+002f, 597 9.77904129e+000f, 2.49660202e+002f, 1.38458405e+001f, 2.14502701e+002f, 598 1.50636337e+002f, 2.15597183e+002f, 6.43103180e+001f, 2.51667648e+002f, 599 1.54952499e+002f, 2.20780014e+002f, 1.26638412e+002f, 2.43040924e+002f, 600 3.67568909e+002f, 1.83624954e+001f, 1.60657944e+002f, 2.21794052e+002f, 601 -1.29507828e+000f, 3.32472443e+002f, 8.51442242e+000f, 4.15561554e+002f, 602 1.27161377e+002f, 1.97260361e+002f, 5.40714645e+000f, 4.90978302e+002f, 603 2.25571690e+001f, 3.96912415e+002f, 2.95664978e+002f, 7.36064959e+000f, 604 1.27241104e+002f, 1.98887573e+002f, -1.25569367e+000f, 3.87713226e+002f, 605 1.04194012e+001f, 4.31495758e+002f, 1.25868874e+002f, 1.99751617e+002f, 606 1.28195480e+002f, 2.02270355e+002f, 2.23436356e+002f, 1.80489182e+002f, 607 1.28727692e+002f, 2.11185410e+002f, 2.03336639e+002f, 2.52182083e+002f, 608 1.29366486e+002f, 2.12201904e+002f, 1.23897598e+002f, 2.17847351e+002f, 609 1.29015259e+002f, 2.19560623e+002f 610 }; 611 612 int npoints = (int)(sizeof(pt1data)/sizeof(pt1data[0])/2); 613 614 Mat p1(1, npoints, CV_32FC2, pt1data); 615 Mat p2(1, npoints, CV_32FC2, pt2data); 616 Mat mask; 617 618 Mat h = findHomography(p1, p2, RANSAC, 0.01, mask); 619 ASSERT_TRUE(!h.empty()); 620 621 transpose(mask, mask); 622 Mat p3, mask2; 623 int ninliers = countNonZero(mask); 624 Mat nmask[] = { mask, mask }; 625 merge(nmask, 2, mask2); 626 perspectiveTransform(p1, p3, h); 627 mask2 = mask2.reshape(1); 628 p2 = p2.reshape(1); 629 p3 = p3.reshape(1); 630 double err = norm(p2, p3, NORM_INF, mask2); 631 632 printf("ninliers: %d, inliers err: %.2g\n", ninliers, err); 633 ASSERT_GE(ninliers, 10); 634 ASSERT_LE(err, 0.01); 635 } 636 637 TEST(Calib3d_Homography, fromImages) 638 { 639 Mat img_1 = imread(cvtest::TS::ptr()->get_data_path() + "cv/optflow/image1.png", 0); 640 Mat img_2 = imread(cvtest::TS::ptr()->get_data_path() + "cv/optflow/image2.png", 0); 641 Ptr<ORB> orb = ORB::create(); 642 vector<KeyPoint> keypoints_1, keypoints_2; 643 Mat descriptors_1, descriptors_2; 644 orb->detectAndCompute( img_1, Mat(), keypoints_1, descriptors_1, false ); 645 orb->detectAndCompute( img_2, Mat(), keypoints_2, descriptors_2, false ); 646 647 //-- Step 3: Matching descriptor vectors using Brute Force matcher 648 BFMatcher matcher(NORM_HAMMING,false); 649 std::vector< DMatch > matches; 650 matcher.match( descriptors_1, descriptors_2, matches ); 651 652 double max_dist = 0; double min_dist = 100; 653 //-- Quick calculation of max and min distances between keypoints 654 for( int i = 0; i < descriptors_1.rows; i++ ) 655 { 656 double dist = matches[i].distance; 657 if( dist < min_dist ) min_dist = dist; 658 if( dist > max_dist ) max_dist = dist; 659 } 660 661 //-- Draw only "good" matches (i.e. whose distance is less than 3*min_dist ) 662 std::vector< DMatch > good_matches; 663 for( int i = 0; i < descriptors_1.rows; i++ ) 664 { 665 if( matches[i].distance <= 100 ) 666 good_matches.push_back( matches[i]); 667 } 668 669 //-- Localize the model 670 std::vector<Point2f> pointframe1; 671 std::vector<Point2f> pointframe2; 672 for( int i = 0; i < (int)good_matches.size(); i++ ) 673 { 674 //-- Get the keypoints from the good matches 675 pointframe1.push_back( keypoints_1[ good_matches[i].queryIdx ].pt ); 676 pointframe2.push_back( keypoints_2[ good_matches[i].trainIdx ].pt ); 677 } 678 679 Mat H0, H1, inliers0, inliers1; 680 double min_t0 = DBL_MAX, min_t1 = DBL_MAX; 681 for( int i = 0; i < 10; i++ ) 682 { 683 double t = (double)getTickCount(); 684 H0 = findHomography( pointframe1, pointframe2, RANSAC, 3.0, inliers0 ); 685 t = (double)getTickCount() - t; 686 min_t0 = std::min(min_t0, t); 687 } 688 int ninliers0 = countNonZero(inliers0); 689 for( int i = 0; i < 10; i++ ) 690 { 691 double t = (double)getTickCount(); 692 H1 = findHomography( pointframe1, pointframe2, RHO, 3.0, inliers1 ); 693 t = (double)getTickCount() - t; 694 min_t1 = std::min(min_t1, t); 695 } 696 int ninliers1 = countNonZero(inliers1); 697 double freq = getTickFrequency(); 698 printf("nfeatures1 = %d, nfeatures2=%d, matches=%d, ninliers(RANSAC)=%d, " 699 "time(RANSAC)=%.2fmsec, ninliers(RHO)=%d, time(RHO)=%.2fmsec\n", 700 (int)keypoints_1.size(), (int)keypoints_2.size(), 701 (int)good_matches.size(), ninliers0, min_t0*1000./freq, ninliers1, min_t1*1000./freq); 702 703 ASSERT_TRUE(!H0.empty()); 704 ASSERT_GE(ninliers0, 80); 705 ASSERT_TRUE(!H1.empty()); 706 ASSERT_GE(ninliers1, 80); 707 } 708
